Fast Determinant of Hessian Filtering for Image Tiepoint Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image tiepoint conjugate matching techniques are computationally expensive and inefficient, especially with large images, due to the 'brute force' approach and failure in featureless areas, leading to wasted processing and poor results.
Innovation Solution
Implementing a hybrid approach that combines fast determinant of Hessian (DOH) filtering for initial feature finding and geometric support data to identify candidate tiepoints, reducing unnecessary processing and improving matching accuracy by focusing on feature-rich areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a brute force approach is used to layout candidate tiepoints, then comprehensive coverage of image areas is achieved, but processing time increases to hours and computational cost becomes excessive
Solution Approach 1:
The patent applies preliminary action by performing determinant of Hessian filtering before the brute force tiepoint matching. This pre-processing step identifies feature-rich areas and eliminates featureless regions in advance, so that subsequent expensive image matching operations are only performed in promising areas, dramatically reducing processing time while maintaining matching accuracy
Solution Approach 2:
The patent extracts and removes featureless image areas from further processing by using determinant of Hessian filtering to identify and eliminate regions with low texture or contrast. This extraction of problematic areas prevents wasted computational resources on matches that would inevitably fail, reducing overall processing time
2Reliability
If brute force layout of candidate grid points is executed, then all image areas are processed, but processing in featureless areas results in wasted computational resources
Solution Approach 1:
The determinant of Hessian filtering is performed as a preliminary action before tiepoint matching to pre-identify feature-rich areas. This allows the system to concentrate computational energy only in areas likely to produce successful matches, eliminating wasted processing in featureless regions
Solution Approach 2:
The patent applies local quality by treating different image areas differently based on their feature content. Feature-rich areas identified by positive determinant of Hessian values receive full processing attention, while featureless areas are eliminated from processing, optimizing computational resource allocation across the image
Data Source
AI summary
A system identifies strong features in conjugate digital images and correlates the conjugate digital images for an identification of candidate tie points using only portions of the conjugate digital images that include the strong features.